Abstract. -Through a series of replacement experiments with the bluehead wrasse, Thalassoma bifasciatum, we have identified male morphological characteristics that appear to be under phenotypic sexual selection. We were particularly interested in whether the various sources of sexual selection (male-male competition for unoccupied mating sites, defense ofmating sites against small males, and female choice of males) were (I) independently associated with different phenotypic characteristics; (2) jointly affected the same characteristic in the same way; or (3) jointly affected the same characteristic in an antagonistic fashion. We replaced the resident large, brightly colored Terminal Phase (TP) males on a reef with the same number ofTP males from other reefs. When transplanted, these males contest with each other to take over mating sites. The transplanted group of males were then scored for three components of fitness: (1) the quality of the site obtained through competition with other large males; (2) the male's ability to defend arriving females from small intruding males; and (3) changes in female visits to the site once the new male takes over. The first and second components are part of intrasexual selection; the third represents intersexual selection. We measured the opportunity for selection by partitioning variance in mating success, and measured the direct effects ofsexual selection by estimating the covariance between morphology and fitness components. Opportunities for selection: Because females generally remain faithful to particular mating sites, most (54%) of the explainable variation in male mating success is due to the acquisition of a particular mating territory, which is the outcome of competition among TP males. There was less variation in mating success due to shifts in site use by females and defense of females against the intrusions of smaller males, but all components were significant. Effects of selection: Success in male-male competition among TP males, estimated by the quality of the territory acquired, was positively associated with body length and the relative length ofthe pectoral fin. Success in territorial defense against small males was primarily related to body length, with lesser contributions from body depth and the area of a white band on the flank. Contribution to fitness through female choice of males was positively associated with white band area. In the two instances where a character was associated with two fitness components, the direction of selection was the same. While body length was positively associated with winning intrasexual contests, it was not correlated to any behavioral measures of aggression. Similarly, the white band associated with attractiveness was not correlated with any aspect of courtship or aggression. Parasite load was uncorrelated with other morphological characters, and did not appear to affect any aspect of sexual selection. There was no evidence for stabilizing selection or significant additional contributions from second-order effects to the ...